Re: URIs and Unique IDs

On Nov 6, 2008, at 2:38 PM, Michael Lang wrote:

> If an ontology is being used by a community to build a description  
> of a domain and accumulate facts about the domain, then the concept  
> of versions is obsolete. The ontology by definition changes over  
> time and becomes more valuable the more it changes.

So this is saying that no one in the community that's building said  
description wants any of the following:
  - a record of when facts were entered and by whom
  - a record of when facts were changed and by whom
  - an ability to compare facts as of today to facts at a historical  
  - an ability to see what has changed that might have caused their  
inference engine to break today when it was working yesterday
  - an ability to easily recognize when they need to re-read the  
ontologies that are being used in their analyses

The ability to see where you were yesterday (and revert to it if  
necessary) is fundamental to software engineering, I thought it would  
be equally fundamental to knowledge engineering (sorry if that isn't  
the appropriate term here -- I am referring to any inferencing using  
distributed resources).

On Nov 6, 2008, at 4:28 PM, Michael F Uschold wrote:
> What this discussion has made clear to me is the existence of two  
> very different kind of use cases and cultures. Those that are Web  
> 2.0 ish, open and dynamic, which may never be used for mission- 
> critical systems and more closed systems

I do not agree that 'open and dynamic' and 'mission critical' are  
mutually exclusive, or perhaps more to the point, that versioning is  
in any way exclusive of 'open and dynamic'.  I cite wikipedia as my  
reference example of versioning in a highly dynamic and open  


John Graybeal   <>  -- 831-775-1956
Monterey Bay Aquarium Research Institute
Marine Metadata Interoperability Project:

Received on Friday, 7 November 2008 00:45:58 UTC